Pub Date : 2026-02-03DOI: 10.1016/j.optcom.2026.132984
Gyaprasad , Rajneesh Joshi
We theoretically investigate a novel mechanism for controlling the quantum degree of polarization of single- and multi-photon light fields through the combined effects of birefringence and dispersion in optical media. While birefringence alone introduces a unitary phase shift between horizontal (H) and vertical (V) polarization modes, the inclusion of dispersion produces frequency-dependent effects that couple polarization with spectral degrees of freedom, thereby inducing decoherence and transforming the quantum state into a mixed state. By employing an electro-optically controlled nematic liquid crystal as the birefringent medium, this decoherence process can be harnessed to achieve tunable control of the quantum degree of polarization. We model this voltage-dependent tunability theoretically and propose methods for experimental verification.
{"title":"Tunable decoherence of quantum polarization states via birefringence-frequency coupling using liquid crystal","authors":"Gyaprasad , Rajneesh Joshi","doi":"10.1016/j.optcom.2026.132984","DOIUrl":"10.1016/j.optcom.2026.132984","url":null,"abstract":"<div><div>We theoretically investigate a novel mechanism for controlling the quantum degree of polarization of single- and multi-photon light fields through the combined effects of birefringence and dispersion in optical media. While birefringence alone introduces a unitary phase shift between horizontal (H) and vertical (V) polarization modes, the inclusion of dispersion produces frequency-dependent effects that couple polarization with spectral degrees of freedom, thereby inducing decoherence and transforming the quantum state into a mixed state. By employing an electro-optically controlled nematic liquid crystal as the birefringent medium, this decoherence process can be harnessed to achieve tunable control of the quantum degree of polarization. We model this voltage-dependent tunability theoretically and propose methods for experimental verification.</div></div>","PeriodicalId":19586,"journal":{"name":"Optics Communications","volume":"608 ","pages":"Article 132984"},"PeriodicalIF":2.5,"publicationDate":"2026-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122640","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-02DOI: 10.1016/j.optcom.2026.132970
Masato Shotoku, Fan Wang, Tomoyoshi Ito, Tomoyoshi Shimobaba
This study proposes a holographic signal converter that transforms existing video equipment such as game consoles, televisions, and computers into holographic displays. This converter transforms the two-dimensional (2D) video signal output from existing video equipment into a three-dimensional (3D) hologram signal without any changes to the video equipment itself. This conversion uses a deep neural network to estimate the depth and a layer method to generate a hologram from the RGB and inferred depth images. Real-time conversion of a 2D video signal to a 3D hologram signal was achieved using the constructed holographic signal converter. We demonstrate a real-time holographic 3D TV game with a PlayStation 5 using our holographic signal converter.
{"title":"Holographic signal converter for existing two-dimensional video equipment","authors":"Masato Shotoku, Fan Wang, Tomoyoshi Ito, Tomoyoshi Shimobaba","doi":"10.1016/j.optcom.2026.132970","DOIUrl":"10.1016/j.optcom.2026.132970","url":null,"abstract":"<div><div>This study proposes a holographic signal converter that transforms existing video equipment such as game consoles, televisions, and computers into holographic displays. This converter transforms the two-dimensional (2D) video signal output from existing video equipment into a three-dimensional (3D) hologram signal without any changes to the video equipment itself. This conversion uses a deep neural network to estimate the depth and a layer method to generate a hologram from the RGB and inferred depth images. Real-time conversion of a 2D video signal to a 3D hologram signal was achieved using the constructed holographic signal converter. We demonstrate a real-time holographic 3D TV game with a PlayStation 5 using our holographic signal converter.</div></div>","PeriodicalId":19586,"journal":{"name":"Optics Communications","volume":"608 ","pages":"Article 132970"},"PeriodicalIF":2.5,"publicationDate":"2026-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122641","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.optcom.2026.132963
Mengmeng Shen , Jiacheng Luo , Fengjuan Chen , Haihua Zhang , Yong Li
It plays a crucial role in virtual reality (VR) and augmented reality (AR) to accurately capture and display real-world three-dimensional (3D) scene. In order to achieve this goal, we propose a new method based on deep-learning to convert deformed fringe in fringe projection profilometry (FPP) to computer-generated hologram (CGH). Three deep neural networks (DNNs) with the same basic structure are designed and trained with three datasets. The intensity information of real scene from a deformed fringe pattern is obtained with the first deep neural network (DNN1) and the phase of deformed fringe is obtained with the second one (DNN2). The hologram is generated from intensity information and phase difference of deformed fringe with the third deep neural network (DNN3). The phase difference is obtained by subtracting reference phase from absolute phase which is obtained with the three-frequency heterodyne algorithm. To improve the efficiency of dataset construction, a new transfer learning approach is proposed. The DNNs are firstly trained with datasets consisting of massive high-fidelity simulated data and then trained with datasets consisting of a small amount of real data. The simulated data are generated using the mathematical model of FPP system and the parameters obtained by system calibration, with ambient illumination, shadows and occlusion considered. Experimental results demonstrate the effectiveness and reliability of the proposed method.
{"title":"Accurate capture and display of real-world three-dimensional scene using fringe projection and deep-learning","authors":"Mengmeng Shen , Jiacheng Luo , Fengjuan Chen , Haihua Zhang , Yong Li","doi":"10.1016/j.optcom.2026.132963","DOIUrl":"10.1016/j.optcom.2026.132963","url":null,"abstract":"<div><div>It plays a crucial role in virtual reality (VR) and augmented reality (AR) to accurately capture and display real-world three-dimensional (3D) scene. In order to achieve this goal, we propose a new method based on deep-learning to convert deformed fringe in fringe projection profilometry (FPP) to computer-generated hologram (CGH). Three deep neural networks (DNNs) with the same basic structure are designed and trained with three datasets. The intensity information of real scene from a deformed fringe pattern is obtained with the first deep neural network (DNN1) and the phase of deformed fringe is obtained with the second one (DNN2). The hologram is generated from intensity information and phase difference of deformed fringe with the third deep neural network (DNN3). The phase difference is obtained by subtracting reference phase from absolute phase which is obtained with the three-frequency heterodyne algorithm. To improve the efficiency of dataset construction, a new transfer learning approach is proposed. The DNNs are firstly trained with datasets consisting of massive high-fidelity simulated data and then trained with datasets consisting of a small amount of real data. The simulated data are generated using the mathematical model of FPP system and the parameters obtained by system calibration, with ambient illumination, shadows and occlusion considered. Experimental results demonstrate the effectiveness and reliability of the proposed method.</div></div>","PeriodicalId":19586,"journal":{"name":"Optics Communications","volume":"607 ","pages":"Article 132963"},"PeriodicalIF":2.5,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080069","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.optcom.2026.132959
Haiming Lu , Geyou Zhang , Rui Gao , Tong Zhou , Bo Zhang , Bin Xu , Kai Liu
One-shot structured light enables real-time 3D scanning, while typically suffering from poor accuracy. In fringe projection profilometry, phase sensitivity is maximized when the phase variation direction is orthogonal to the epipolar lines of the projector. Straight fringes are quasi-optimal under certain conditions. However, they are sensitive to system setup, and misalignment can cause noticeable accuracy loss. Circular fringe patterns overcome this limitation by achieving globally optimal phase sensitivity. In this paper, we present a generalized framework for one-shot FPP using optimal circular fringes. First, we develop a filter based on geometric analysis of the spectrum to extract the wrapped phase. Second, phase unwrapping and mapping are conducted to restore absolute phase. Finally we reconstruct 3D points via the extended epipolar geometry. Experiments show that circular fringes significantly improve reconstruction accuracy, especially in fine details, highlighting their superiority over straight fringes.
{"title":"One-shot optimal circular fringe projection profilometry","authors":"Haiming Lu , Geyou Zhang , Rui Gao , Tong Zhou , Bo Zhang , Bin Xu , Kai Liu","doi":"10.1016/j.optcom.2026.132959","DOIUrl":"10.1016/j.optcom.2026.132959","url":null,"abstract":"<div><div>One-shot structured light enables real-time 3D scanning, while typically suffering from poor accuracy. In fringe projection profilometry, phase sensitivity is maximized when the phase variation direction is orthogonal to the epipolar lines of the projector. Straight fringes are quasi-optimal under certain conditions. However, they are sensitive to system setup, and misalignment can cause noticeable accuracy loss. Circular fringe patterns overcome this limitation by achieving globally optimal phase sensitivity. In this paper, we present a generalized framework for one-shot FPP using optimal circular fringes. First, we develop a filter based on geometric analysis of the spectrum to extract the wrapped phase. Second, phase unwrapping and mapping are conducted to restore absolute phase. Finally we reconstruct 3D points via the extended epipolar geometry. Experiments show that circular fringes significantly improve reconstruction accuracy, especially in fine details, highlighting their superiority over straight fringes.</div></div>","PeriodicalId":19586,"journal":{"name":"Optics Communications","volume":"608 ","pages":"Article 132959"},"PeriodicalIF":2.5,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122639","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Compared to traditional sensors, metasurface sensors offer higher sensitivity and superior optical response. All-dielectric materials have broad prospects for refractive index sensing due to low ohmic loss. Here, we propose a double-rod all-dielectric nanostructure that exhibits a high-quality factor (high-Q) Fano resonance in the mid-infrared band. This resonance is highly sensitive to changes in the refractive index of the surrounding medium. Analysis of the electromagnetic field distribution and multipole moment decomposition, it is demonstrated that the resonance is driven by a toroidal dipole (TD) and a magnetic quadrupole (MQ). We systematically characterized the sensing performance of the proposed structure. The results show that it achieves a sensitivity of up to 1337.1 nm/RIU and a high figure of merit (FOM) of 1238. In the mid-infrared band, the structure exhibited a high Q factor of 18544. Moreover, its reflection spectrum in this band could be effectively tuned by adjusting the geometric parameters of the metasurface. Finally, investigations at different incident angles reveal that the resonant peak exhibits a distinct blueshift as the angle increases. Moreover, the structure shows a selective response to the polarization state, demonstrating excellent polarization sensitivity. This work shows that high-performance optical sensors can be fabricated using simple processes, thereby providing a fresh design framework and theoretical basis for the sensor community.
{"title":"High-Q mid-infrared refractive index sensor based on Fano resonance in an all-dielectric double-rod structure","authors":"Wenwen Wang, Fuming Yang, Wenwen Sun, Zhe Wu, Xiaoyan Shi, Junying Liu, Yuetao Liu, Jizheng Geng, Xintong Wei, Xiangtao Chen, Shijia Zhu, Zhongzhu Liang","doi":"10.1016/j.optcom.2026.132975","DOIUrl":"10.1016/j.optcom.2026.132975","url":null,"abstract":"<div><div>Compared to traditional sensors, metasurface sensors offer higher sensitivity and superior optical response. All-dielectric materials have broad prospects for refractive index sensing due to low ohmic loss. Here, we propose a double-rod all-dielectric nanostructure that exhibits a high-quality factor (high-Q) Fano resonance in the mid-infrared band. This resonance is highly sensitive to changes in the refractive index of the surrounding medium. Analysis of the electromagnetic field distribution and multipole moment decomposition, it is demonstrated that the resonance is driven by a toroidal dipole (TD) and a magnetic quadrupole (MQ). We systematically characterized the sensing performance of the proposed structure. The results show that it achieves a sensitivity of up to 1337.1 nm/RIU and a high figure of merit (FOM) of 1238. In the mid-infrared band, the structure exhibited a high Q factor of 18544. Moreover, its reflection spectrum in this band could be effectively tuned by adjusting the geometric parameters of the metasurface. Finally, investigations at different incident angles reveal that the resonant peak exhibits a distinct blueshift as the angle increases. Moreover, the structure shows a selective response to the polarization state, demonstrating excellent polarization sensitivity. This work shows that high-performance optical sensors can be fabricated using simple processes, thereby providing a fresh design framework and theoretical basis for the sensor community.</div></div>","PeriodicalId":19586,"journal":{"name":"Optics Communications","volume":"608 ","pages":"Article 132975"},"PeriodicalIF":2.5,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122642","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-27DOI: 10.1016/j.optcom.2026.132962
Chenlong Dai, Yang Wang, Jiantai Dou
Accurate axial distance calibration is pivotal for high-precision ptychography. The conventional dichotomy-based approach (dPIE) suffers from slow convergence and computationally intensive preprocessing. Here, we introduce a hybrid framework: A linear model analytically reconstructs high-fidelity initial object/probe distributions, integrated with an optimized dichotomous scheme for axial refinement. This strategy enables single-iteration processing at each candidate distance without preprocessing overhead, which effectively overcomes key limitations of dPIE. Validated on standard and biological specimens, the framework can drastically reduce the correction time. By leveraging total variation to exponentially narrow the search interval, the rapid convergence is achieved with minimal computation. This method provides an efficient solution for axial distance calibration in ptychographic systems.
{"title":"Fast calibration method of axial distance error in ptychography based on linear model for initial object reconstruction","authors":"Chenlong Dai, Yang Wang, Jiantai Dou","doi":"10.1016/j.optcom.2026.132962","DOIUrl":"10.1016/j.optcom.2026.132962","url":null,"abstract":"<div><div>Accurate axial distance calibration is pivotal for high-precision ptychography. The conventional dichotomy-based approach (dPIE) suffers from slow convergence and computationally intensive preprocessing. Here, we introduce a hybrid framework: A linear model analytically reconstructs high-fidelity initial object/probe distributions, integrated with an optimized dichotomous scheme for axial refinement. This strategy enables single-iteration processing at each candidate distance without preprocessing overhead, which effectively overcomes key limitations of dPIE. Validated on standard and biological specimens, the framework can drastically reduce the correction time. By leveraging total variation to exponentially narrow the search interval, the rapid convergence is achieved with minimal computation. This method provides an efficient solution for axial distance calibration in ptychographic systems.</div></div>","PeriodicalId":19586,"journal":{"name":"Optics Communications","volume":"607 ","pages":"Article 132962"},"PeriodicalIF":2.5,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080066","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-27DOI: 10.1016/j.optcom.2025.132805
Xiang Sui , Ying Shang , Sheng Huang , Wenan Zhao , Xiaohan Qiao , Guangqiang Liu , Chunmei Yao , Shouling Liu , Na Wan , Xianggui Kong , Hong Zhao , Fengming Mou , Zhengying Li , Weitao Wang , Chen Wang , Gangding Peng
{"title":"Corrigendum to “Coherent fading suppression method in the COTDR system based on multi-band filtering” [Opt. Commun. 583 (2025) 131696]","authors":"Xiang Sui , Ying Shang , Sheng Huang , Wenan Zhao , Xiaohan Qiao , Guangqiang Liu , Chunmei Yao , Shouling Liu , Na Wan , Xianggui Kong , Hong Zhao , Fengming Mou , Zhengying Li , Weitao Wang , Chen Wang , Gangding Peng","doi":"10.1016/j.optcom.2025.132805","DOIUrl":"10.1016/j.optcom.2025.132805","url":null,"abstract":"","PeriodicalId":19586,"journal":{"name":"Optics Communications","volume":"607 ","pages":"Article 132805"},"PeriodicalIF":2.5,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080154","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-27DOI: 10.1016/j.optcom.2026.132966
Maoni Chen , Aimin Liu , Shangde Zhou , Jianqiu Cao , Qi Zhang , Zhihe Huang , Zilun Chen , Zefeng Wang , Jinbao Chen
A 1.5-kW large-core step-index Yb-doped fiber amplifier operating near 980 nm is firstly demonstrated, to the best of our knowledge, by means of broadening the signal bandwidth to suppress in-band amplified spontaneous emission (ASE) around 980 nm. The signal bandwidth covering from 974 nm to 980 nm is achieved using a home-made seed oscillator with a dual low-reflectivity fiber Bragg grating (LR-FBG) configuration. The record 1.546-kW output power was achieved with a slope efficiency of 71.5 % (with respect to the launched pump power) which should also be the highest achieved by the step-index double-cladding Yb-doped fiber. The power ratio of in-band ASE was suppressed to below 1 %, while the ASE around 1030 nm was also 27.8-dB suppressed. This amplifier can strongly drive the power up-scaling of high-power fiber lasers and amplifiers as high-power cladding-pumping source, and thus can have significant impact on application fields involving high-power fiber lasers and amplifiers.
{"title":"Demonstration of 1.5-kW large-core step-index Yb-doped fiber amplifier near 980 nm","authors":"Maoni Chen , Aimin Liu , Shangde Zhou , Jianqiu Cao , Qi Zhang , Zhihe Huang , Zilun Chen , Zefeng Wang , Jinbao Chen","doi":"10.1016/j.optcom.2026.132966","DOIUrl":"10.1016/j.optcom.2026.132966","url":null,"abstract":"<div><div>A 1.5-kW large-core step-index Yb-doped fiber amplifier operating near 980 nm is firstly demonstrated, to the best of our knowledge, by means of broadening the signal bandwidth to suppress in-band amplified spontaneous emission (ASE) around 980 nm. The signal bandwidth covering from 974 nm to 980 nm is achieved using a home-made seed oscillator with a dual low-reflectivity fiber Bragg grating (LR-FBG) configuration. The record 1.546-kW output power was achieved with a slope efficiency of 71.5 % (with respect to the launched pump power) which should also be the highest achieved by the step-index double-cladding Yb-doped fiber. The power ratio of in-band ASE was suppressed to below 1 %, while the ASE around 1030 nm was also 27.8-dB suppressed. This amplifier can strongly drive the power up-scaling of high-power fiber lasers and amplifiers as high-power cladding-pumping source, and thus can have significant impact on application fields involving high-power fiber lasers and amplifiers.</div></div>","PeriodicalId":19586,"journal":{"name":"Optics Communications","volume":"607 ","pages":"Article 132966"},"PeriodicalIF":2.5,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-27DOI: 10.1016/j.optcom.2025.132779
Xiang Sui , Ying Shang , Sheng Huang , Wenan Zhao , Xiaohan Qiao , Guangqiang Liu , Chunmei Yao , Shouling Liu , Na Wan , Xianggui Kong , Hong Zhao , Fengming Mou , Zhengying Li , Weitao Wang , Chen Wang , Gangding Peng
{"title":"Corrigendum to “Coherent fading suppression method in the COTDR system based on multi-band filtering” [Opt. Commun. 583 (2025) 131696]","authors":"Xiang Sui , Ying Shang , Sheng Huang , Wenan Zhao , Xiaohan Qiao , Guangqiang Liu , Chunmei Yao , Shouling Liu , Na Wan , Xianggui Kong , Hong Zhao , Fengming Mou , Zhengying Li , Weitao Wang , Chen Wang , Gangding Peng","doi":"10.1016/j.optcom.2025.132779","DOIUrl":"10.1016/j.optcom.2025.132779","url":null,"abstract":"","PeriodicalId":19586,"journal":{"name":"Optics Communications","volume":"607 ","pages":"Article 132779"},"PeriodicalIF":2.5,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080153","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-24DOI: 10.1016/j.optcom.2026.132958
Cui Zhang , Xinjun Gao , Ajiaikebaier Wulamu , Weibing Gan , Ai Zhou , Congcong Qin , Pei Lv
Natural gas pipelines are critical infrastructure for urban energy supply, but leak-induced vibration signals are often obscured by environmental noise and coupling from supporting structures, which reduces the accuracy of detection and localization. This paper proposes a denoising framework that integrates a weak fiber Bragg grating (WFBG) array with deep learning. Specifically, the method builds a VMD-fused multi-branch structure, in which U-Net and gated recurrent unit (GRU) subnetworks are used for feature extraction and are trained under the Noise2Noise (N2N) framework, thereby enabling denoising without clean reference signals. It effectively achieves extraction and reconstruction of both high- and low-frequency features. Experiments are conducted on a steel pipeline leakage monitoring platform with synchronous multi-zone acquisition (Zone 1 to Zone 9). Training pairs are constructed using data from different time instants within the same zone, and cross-zone consistency is validated under different noise and coupling conditions. For physical interpretability, this paper provides an order-of-magnitude prediction of the first bending mode of a supported span to explain the dominant low-frequency peak and its spatial variation. Under representative operating conditions, the proposed method achieves , , , time-domain , and frequency-domain error , outperforming baseline methods. Overall, the framework preserves key frequency-band characteristics related to leakage while suppressing noise, and demonstrates robust cross-zone generalization for distributed monitoring.
{"title":"Vibration signal denoising method for pipeline leakage based on WFBG and deep learning","authors":"Cui Zhang , Xinjun Gao , Ajiaikebaier Wulamu , Weibing Gan , Ai Zhou , Congcong Qin , Pei Lv","doi":"10.1016/j.optcom.2026.132958","DOIUrl":"10.1016/j.optcom.2026.132958","url":null,"abstract":"<div><div>Natural gas pipelines are critical infrastructure for urban energy supply, but leak-induced vibration signals are often obscured by environmental noise and coupling from supporting structures, which reduces the accuracy of detection and localization. This paper proposes a denoising framework that integrates a weak fiber Bragg grating (WFBG) array with deep learning. Specifically, the method builds a VMD-fused multi-branch structure, in which U-Net and gated recurrent unit (GRU) subnetworks are used for feature extraction and are trained under the Noise2Noise (N2N) framework, thereby enabling denoising without clean reference signals. It effectively achieves extraction and reconstruction of both high- and low-frequency features. Experiments are conducted on a steel pipeline leakage monitoring platform with synchronous multi-zone acquisition (Zone 1 to Zone 9). Training pairs are constructed using data from different time instants within the same zone, and cross-zone consistency is validated under different noise and coupling conditions. For physical interpretability, this paper provides an order-of-magnitude prediction of the first bending mode of a supported span to explain the dominant low-frequency peak and its spatial variation. Under representative operating conditions, the proposed method achieves <span><math><mrow><mi>SNR</mi><mo>=</mo><mn>20</mn><mo>.</mo><mn>32</mn><mspace></mspace><mi>dB</mi></mrow></math></span>, <span><math><mrow><mi>PSNR</mi><mo>=</mo><mn>31</mn><mo>.</mo><mn>71</mn><mspace></mspace><mi>dB</mi></mrow></math></span>, <span><math><mrow><mi>SSIM</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>929</mn></mrow></math></span>, time-domain <span><math><mrow><mi>MSE</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>00581</mn></mrow></math></span>, and frequency-domain error <span><math><mrow><msub><mrow><mi>Err</mi></mrow><mrow><mi>f</mi></mrow></msub><mo>=</mo><mn>1</mn><mo>.</mo><mn>99985</mn></mrow></math></span>, outperforming baseline methods. Overall, the framework preserves key frequency-band characteristics related to leakage while suppressing noise, and demonstrates robust cross-zone generalization for distributed monitoring.</div></div>","PeriodicalId":19586,"journal":{"name":"Optics Communications","volume":"607 ","pages":"Article 132958"},"PeriodicalIF":2.5,"publicationDate":"2026-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080152","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号